I agree-- a double-blind selection method is best, with overall results made visible. I'm thinking the process needs to be staged in an environment outside the wiki and published there with all results afterward. Of course, that leaves open the question for what staging vehicle...